PhD project: Neural decoding of motor cortex dynamics during learning of skilled arm movements

State of the art 

A fundamental problem in modern neuroscience is to understand how the brain generates the rich repertoire of movements that humans and other animals exhibit. Decades of experimental work have highlighted the role of the motor cortex in issuing the high-level motor commands that are send down the spinal cord to activate muscles. Yet, the underlying “neural code”, that is, the operation by which dynamic patterns of neural activity are transformed into the associated kinematic parameters is still not fully characterized. In particular, it remains unknown whether a “universal decoder” can be used to translate ongoing neural activity into features of movement across a wide range of behavioral contexts. 

Objectives 

This PhD project aims at exploring the relationship between motor cortical activity and upper-limb movements. One main objective is to investigate the role of neural activity preceding movement, also called preparatory activity, in determining the properties of the upcoming movement. We hypothesize that the contribution of preparatory activity to movement execution evolves with practice, as the motor cortex goes from being externally-driven to an autonomous movement generator.  

Methods 

To test this hypothesis, the successful candidate will have access to a rich behavioral and neural dataset collected in two macaque monkeys performing complex hand reaching movements. Movements were recorded in the context of an “unconstrained” motor learning task, thus providing a unique opportunity to characterize the properties of motor cortical activity over extensive practice periods and across a wide range of movements. The student will develop and test algorithms that will attempt to predict with high precision the instantaneous hand movement trajectories of the animals based on their ongoing neural activity. These algorithms will be guided by theoretical considerations, including the hypothesized role of preparatory activity in determining movement characteristics.  

Expected results 

The project will offer exciting prospects for the candidate to explore both basic and applied aspects of neuroscience research. From a fundamental science perspective, the project will provide important insights into the computational principles supporting the generation of complex movements. From an applied sciences perspective, the goals of the project are aligned with the growing interests in the neuro-engineering community to define robust decoding algorithms that could ultimately be used to drive neuro-prosthetics in paralyzed patients. 

Feasibility  

The dataset used in the project has already been collected (two monkeys) which minimizes the experimental risks/delays and will help expedite the publication process. Moreover, the student will be guided in their work by an expert in monkey electrophysiology and motor control (Thomas Brochier) and assisted by a senior postdoctoral fellow (Nicolas Meirhaeghe) in all aspects of the quantitative analyses. 

Expected candidate profile  

The ideal candidate should have an inclination for brain sciences and sufficient background in mathematics (linear algebra) and computer science to handle complex data (fluent in Python/Matlab). Interests or prior experience in dynamical systems, machine learning, and brain-computer interfaces would be a plus.